Differential Privacy in Cooperative Multiagent Planning

Bo Chen, Calvin Hawkins, Mustafa O. Karabag, Cyrus Neary, Matthew Hale, Ufuk Topcu
Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence, PMLR 216:347-357, 2023.

Abstract

Privacy-aware multiagent systems must protect agents’ sensitive data while simultaneously ensuring that agents accomplish their shared objectives. Towards this goal, we propose a framework to privatize inter-agent communications in cooperative multiagent decision-making problems. We study sequential decision-making problems formulated as cooperative Markov games with reach-avoid objectives. We apply a differential privacy mechanism to privatize agents’ communicated symbolic state trajectories, and analyze tradeoffs between the strength of privacy and the team’s performance. For a given level of privacy, this tradeoff is shown to depend critically upon the total correlation among agents’ state-action processes. We synthesize policies that are robust to privacy by reducing the value of the total correlation. Numerical experiments demonstrate that the team’s performance under these policies decreases by only 6 percent when comparing private versus non-private implementations of communication. By contrast, the team’s performance decreases by 88 percent when using baseline policies that ignore total correlation and only optimize team performance.

Cite this Paper

BibTeX
@InProceedings{pmlr-v216-chen23e, title = {Differential Privacy in Cooperative Multiagent Planning}, author = {Chen, Bo and Hawkins, Calvin and Karabag, Mustafa O. and Neary, Cyrus and Hale, Matthew and Topcu, Ufuk}, booktitle = {Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence}, pages = {347--357}, year = {2023}, editor = {Evans, Robin J. and Shpitser, Ilya}, volume = {216}, series = {Proceedings of Machine Learning Research}, month = {31 Jul--04 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v216/chen23e/chen23e.pdf}, url = {https://proceedings.mlr.press/v216/chen23e.html}, abstract = {Privacy-aware multiagent systems must protect agents’ sensitive data while simultaneously ensuring that agents accomplish their shared objectives. Towards this goal, we propose a framework to privatize inter-agent communications in cooperative multiagent decision-making problems. We study sequential decision-making problems formulated as cooperative Markov games with reach-avoid objectives. We apply a differential privacy mechanism to privatize agents’ communicated symbolic state trajectories, and analyze tradeoffs between the strength of privacy and the team’s performance. For a given level of privacy, this tradeoff is shown to depend critically upon the total correlation among agents’ state-action processes. We synthesize policies that are robust to privacy by reducing the value of the total correlation. Numerical experiments demonstrate that the team’s performance under these policies decreases by only 6 percent when comparing private versus non-private implementations of communication. By contrast, the team’s performance decreases by 88 percent when using baseline policies that ignore total correlation and only optimize team performance.} }
Endnote
%0 Conference Paper %T Differential Privacy in Cooperative Multiagent Planning %A Bo Chen %A Calvin Hawkins %A Mustafa O. Karabag %A Cyrus Neary %A Matthew Hale %A Ufuk Topcu %B Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2023 %E Robin J. Evans %E Ilya Shpitser %F pmlr-v216-chen23e %I PMLR %P 347--357 %U https://proceedings.mlr.press/v216/chen23e.html %V 216 %X Privacy-aware multiagent systems must protect agents’ sensitive data while simultaneously ensuring that agents accomplish their shared objectives. Towards this goal, we propose a framework to privatize inter-agent communications in cooperative multiagent decision-making problems. We study sequential decision-making problems formulated as cooperative Markov games with reach-avoid objectives. We apply a differential privacy mechanism to privatize agents’ communicated symbolic state trajectories, and analyze tradeoffs between the strength of privacy and the team’s performance. For a given level of privacy, this tradeoff is shown to depend critically upon the total correlation among agents’ state-action processes. We synthesize policies that are robust to privacy by reducing the value of the total correlation. Numerical experiments demonstrate that the team’s performance under these policies decreases by only 6 percent when comparing private versus non-private implementations of communication. By contrast, the team’s performance decreases by 88 percent when using baseline policies that ignore total correlation and only optimize team performance.
APA
Chen, B., Hawkins, C., Karabag, M.O., Neary, C., Hale, M. & Topcu, U.. (2023). Differential Privacy in Cooperative Multiagent Planning. Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 216:347-357 Available from https://proceedings.mlr.press/v216/chen23e.html.

Related Material